58 BULLETIN 903, U. S. DEPARTMENT OF AGEICULTUEE. 



occasionally that some of the eggs used in the inoculations dropped 

 off. This was unavoidable, and when egg-laying females were 

 under observation it frequently happened that the eggs dropped 

 down. When more than one female was producing eggs simul- 

 taneously on -a single root, there even arose doubt as to which certain 

 of the fallen eggs should be credited. For the biological records 

 the first season, 9 vines were used, averaging 3 separate roots each, 

 but since 3 of these roots died, after they were planted, only 24 

 roots were actually used. Of these 3 -were used for rearing gallicole 

 progeny and 5 others were used for nymph production and ferti- 

 lizer experiments on heavily infested vines, leaving 16 for individual 

 records. In 1914 and 1915, only 4 vines were used each year to con- 

 tinue the individual series. 



Many interesting habits were observed, but the behavior of the 

 phylloxera? in the main did not differ from that observed under 

 cellar conditions upon severed roots. Newly hatched larvae mostly 

 settled close to the eggshells they had vacated, but if there were 

 any fresh lesions near by, the young larva? often found their way 

 to them and settled. Occasional movements of older individuals 

 were observed, not only at the time of molting but at any time in 

 the instars. These movements were generally in the direction of 

 more succulent food. Occasionally egg-laying individuals shifted 

 their positions without apparent injury, although this was some- 

 times followed by a temporary halt in the production of eggs. The 

 production of nymphs occurred from June to October, as in the vine- 

 yard. The tendency of the nymphs to crawl up the root just before 

 transforming was noticeable. Most of them could go no farther up- 

 ward than the cylinder plugged with cotton, and so perforce had 

 to transform into migrants at this point. A small percentage trans- 

 formed at points farther down the root and did not appear to have 

 made any upward migration. On the heavily infested roots, wan- 

 derers appeared from July to October. These roamed around the 

 inside of the compartment, and succeeded in finding their way down 

 through cracks in the soil of the lower pots, and infested the root- 

 lets, especially those growing against the inside of the pots. After 

 irrigation, cracks appeared in the soil around the inside of the pot, 

 furnishing the wandering larva? access to the rootlets. In no case 

 was this infestation of any great extent, although large numbers of 

 wandering larvae were observed in several of the cages, and only a 

 very small percentage, presumably, found their way to a new food 

 supply. This fact has an important bearing upon the distribution 

 of the insect as will be noted elsewhere. Although it appeared dif- 

 ficult for the insects to penetrate an inch of sand in the lower glass 

 cylinder, the occurrence of large infestations immediately below the 

 cylinder, coupled with evidences of only slight infestations on the 



